Twitter Sentiment Analysis through Bidirectional LSTM

The experiment consists of developing a model using Bidirectional LSTM layers in order to predict sentiment of a given dataset.

TensorFlow, Keras, Scikit-Learn was used for developing the relatively simple NLP model in this scenario. The code itself is explained on the notebook file.

For the best view of all Markdown parts and running of code, Google Colaboratory is recommended. The notebook itself wa run on Google Colaboratory under the runtime type GPU. Each epoch approximately took 40 sec to 60 sec to run on an average on GPU runtime.